So, after having this box running outside for over 6 weeks now I get nice results on bright nights, ID a satellite or 2, and add to the seesat community list.
However, getting a bit dissattisfied that about half of the observations show in red text meaning that observation has no / not enough valid stars to confirm the fov of the camera.
Checking several options to improve, the best one is to switch to a monochrom Asi camera wih large sensor and not to high resolution, But that takes some time to decide.
So, in the mean time I checked if there where ways to reduce noise in the images, making more stars clearly visible, so I decided to take a small dive into the “analog gain settings” for the PiHQ camera.
From the RPI forum:
a hq (imx477) camera has a max Analog gain of 22.26, so above that digital gain = gain/ 22.26
So using the default value “90” in configuration.ini sets the camera for full analog gain and 4 times digital gain.
Last night I decided to run it with 2 times digital gain, so I did set it for “45”
The results look promising. I have much more black text observations and only a few in red instead of 30 / 70 black / red.
However, the camera ran in some state of timeout before midnight UTC
[821:47:21.792153824] [1451534] WARN V4L2 v4l2_videodevice.cpp:2096 /dev/video4[25:cap]: Dequeue timer of 1000000.00us has expired!
[821:47:21.949793817] [1451534] ERROR RPI pipeline_base.cpp:1371 Camera frontend has timed out!
So I changed the exposure time from 100000 us to 90000 us and I did change an environmet setting for the timeout
bz-stvid@raspberrypi:~ $ echo $LIBCAMERA_RPI_CONFIG_PARAMS
camera_timeout_value_ms=10000
So, ready for the next night.
I still had 12 satellites in les than 2 hours of obs, from which Yaogan 16 A, B and C within 10 minutes from each other.
Also sent the data for 42072 to the list, my valid obs where off from the catalog position in the 2 obs.
Hope this is helpfull for anyone 